MOCVD為磊晶製程的重要設備，其利用高溫熱裂解的方式將有機金屬源與氨氣等斷鍵，使離子結合並堆積在基板上，與基板產生化學作用並吸附堆積形成薄膜，分子吸附堆積之情形受基板溫度的影響，藉而影響薄膜品質，使其光電元件效能、壽命有優劣之分。而溫度均勻性直接影響磊晶薄膜之品質，因此，本研究以自行搭建之高溫實驗腔體進行以數值分析法設計加熱系統之熱擋板，藉由數值分析之結果再與升溫實驗進行驗證。下列為針對之參數分析 （1） 加熱系統對載盤盤面溫度分佈 （2） 加熱器與熱擋板對於載盤盤面之溫度影響 （3） 載盤厚度差異對盤面溫度分佈之影響 以數值分析結果得知盤面溫度分佈，晶圓區平均溫度差為4.35oC，最大標準差為1.53。針對載盤盤面晶圓區溫度之分佈情形，在載盤厚度減小時，載盤整體溫度差增加，因此厚度10mm載盤之溫度均勻性較其他優良。比較數值分析結果與升溫實驗結果，將兩層側邊熱擋板增設置自行建立之高溫腔體中，進行升溫實驗比對，載盤盤面晶圓區位置之溫度誤差最大值1.2 %，最小值小於0.1 %，由數值分析可知改善溫晶圓區之U型溫度分佈，實際升溫實驗後得知結果是可以利用數值分析法被預期的。 ;Metal Organic Chemical Vapor Deposition (MOCVD) is used for growing the epitaxy thin film, which usually can be divided into five sub-systems, (a) injecting system, (b) vacuum chamber, (c) heating system, (d) exhausting system and (e) control system. The processing condition of a MOCVD requires a high temperature to deposit epitaxy thin film, and the target temperature is higher than 1200oC at the heating system. So the temperature uniformity of the wafer area on susceptor is an important key for qualifying and performance of thin film process. As a result, this research focuses on the heating system of a MOCVD vacuum reactor. The numerical analysis is used to simulate the temperature distribution on the surface of susceptor in a very-high temperature vacuum reactor. The heating baffle is added to heating system not only below the heater but also around the susceptor and heater. This very-high temperature vacuum reactor is a home-made reactor in which size is six pieces of two-inch wafers can be placed, and the thermocouple on the surface of susceptor is used for the temperature measurement. The heating baffle affects the thermal radiation distribution, and the temperature uniformity is progressive on the surface of susceptor. In the numerical analysis result, the temperature difference is 4.35oC at the wafer, and the standard deviation maximum is 1.53 while the heating baffle is added to the very-high temperature vacuum reactor and the thickness of susceptor is 10mm. When the two-tier heating baffle is added to heating system, to compare the results in simulation with experiments, the maximum 1.2% and the minimum 0.1% can be obtained for temperature difference of the wafer on the susceptor surface.
